<p>Efficient global dynamic optimization (GDO) using spatial branch-and-bound (SBB) requires the ability to construct tight bounds for the dynamic model. This thesis works toward efficient GDO by developing effective convex relaxation techniques for models with ordinary differential equations (ODEs). In particular, a novel algorithm, based upon a verified interval ODE method and the McCormick relaxation technique, is developed for constructing convex and concave relaxations of solutions of nonlinear parametric ODEs. In addition to better convergence properties, the relaxations so obtained are guaranteed to be no looser than their underlying interval bounds, and are typically tighter in practice. Moreover, they are rigorous in the sense of accounting for truncation errors. Nonetheless, the tightness of the relaxations is affected by the overestimation from the dependency problem of interval arithmetic that is not addressed systematically in the underlying interval ODE method. To handle this issue, the relaxation algorithm is extended to a Taylor model ODE method, which can provide generally tighter enclosures with better convergence properties than the interval ODE method. This way, an improved version of the algorithm is achieved where the relaxations are generally tighter than those computed with the interval ODE method, and offer better convergence. Moreover, they are guaranteed to be no looser than the interval bounds obtained from Taylor models, and are usually tighter in practice. However, the nonlinearity and (potentially) nonsmoothness of the relaxations impedes their fast and reliable solution. Therefore, the algorithm is finally modified by incorporating polyhedral relaxations in order to generate relatively tight and computationally cheap linear relaxations for the dynamic model. The resulting relaxation algorithm along with a SBB procedure is implemented in the MC++ software package. GDO utilizing the proposed relaxation algorithm is demonstrated to have significantly reduced computational expense, up to orders of magnitude, compared to existing GDO methods.</p> / Doctor of Philosophy (PhD)

Fractional calculus has been used as a research tool in the fields of pharmacology, biology, chemistry, and other areas [3]. The main purpose of this thesis is to calculate Green's functions of fractional difference equations, and to model problems in pharmacokinetics. We claim that the discrete fractional calculus yields the best prediction performance compared to the continuous fractional calculus in the application of a one-compartmental model of drug concentration. In Chapter 1, the Gamma function and its properties are discussed to establish a theoretical basis. Additionally, the basics of discrete fractional calculus are discussed using particular examples for further calculations. In Chapter 2, we use these basic results in the analysis of a linear fractional difference equation. Existence of solutions to this difference equation is then established for both initial conditions (IVP) and two-point boundary conditions (BVP). In Chapter 3, Green's functions are introduced and discussed, along with examples. Instead of using Cauchy functions, the technique of finding Green's functions by a traditional method is demonstrated and used throughout this chapter. The solutions of the BVP play an important role in analysis and construction of the Green's functions. Then, Green's functions for the discrete calculus case are calculated using particular problems, such as boundary value problems, discrete boundary value problems (DBVP) and fractional boundary value problems (FBVP). Finally, we demonstrate how the Green's functions of the FBVP generalize the existence results of the Green's functions of DVBP. In Chapter 4, different compartmental pharmacokinetic models are discussed. This thesis limits discussion to the one-compartmental model. The Mathematica FindFit command and the statistical computational techniques of mean square error (MSE) and cross-validation are discussed. Each of the four models (continuous, continuous fractional, discrete and discrete fractional) is used to compute the MSE numerically with the given data of drug concentration. Then, the best fit and the best model are obtained by inspection of the resulting MSE. In the last Chapter, the results are summarized, conclusions are drawn, and directions for future work are stated.

Ordinary differential equations (ODEs) play an important role in mathematics. Although intrinsically, the setting for describing ODEs is the continuous framework, where differential operators are considered as maps from one function space to another, common numerical algorithms for ODEs discretise problems early on in the solution process. This thesis is about continuous analogues of such discrete algorithms for the numerical solution of ODEs. This thesis shows how Newton's method for finite dimensional system can be generalised to function spaces, where it is known as Newton-Kantorovich iteration. It presents affine invariant damping strategies for increasing the chance of convergence for the Newton-Kantorovich iteration. The derivatives required in this continuous setting are Fr&eacute;chet derivatives, the continuous analogue of Jacobian matrices. In this work, we present how automatic differentiation techniques can be applied to compute Fr&eacute;chet derivatives. We introduce chebop, a Matlab solver for nonlinear boundary-value problems, which combines damped Newton iteration in function space and automatic Fr&eacute;chet differentiation. By proving that affine operators have constant Fr&eacute;chet derivatives, it is demonstrated how automatic linearity detection of computed quantities can be implemented. This is valuable for black-box solvers, which can use the information to determine whether an iteration scheme has to be employed for solving a problem. Like nonlinear systems of equations, nonlinear boundary-value problems can have multiple solutions. This thesis present two techniques for obtaining multiple solutions of operator equations: deflation and path-following. An algorithm combining the two techniques is proposed.

Coupled oscillators, such as groups of fireflies or clusters of neurons, are found throughout nature and are frequently modeled in the applied mathematics literature. Earlier work by Kuramoto, Strogatz, and others has led to a deep understanding of the emergent behavior of systems of such oscillators using traditional dynamical systems methods. In this project we outline the application of techniques from topological data analysis to understanding the dynamics of systems of coupled oscillators. This includes the examination of partitions, partial synchronization, and attractors. By looking for clustering in a data space consisting of the phase change of oscillators over a set of time delays we hope to reconstruct attractors and identify members of these clusters.

The birth of complexity research derives from the logical progression of advancement in the scientific field afforded by reductionist theory. We present in silico models of two complex physiological processes, wound healing and coagulation/fibrinolysis based on two common tools in the study of complex physiology: ordinary differential equations (ODE) and Agent Based Modeling (ABM). The strengths of these two approaches are well-suited in the analysis of clinical paradigms such as wound healing and coagulation. The complex interactions that characterize acute wound healing have stymied the development of effective therapeutic modalities. The use of computational models holds the promise to improve our basic approach to understanding the process. We have modified an existing ordinary differential equation model by 1) evolving from a systemic model to a local model, 2) the incorporation of fibroblast activity, and3) including the effects of tissue oxygenation. Possible therapeutic targets, such as fibroblast death rate and rate of fibroblast recruitment have been identified by computational analysis. This model is a step toward constructing an integrative systems biology model of human wound healing. The coagulation and fibrinolytic systems are complex, inter-connected biological systems with major physiological roles. We present an Agent Based Modeling and Simulation (ABMS) approach to these complex interactions. This ABMS method successfully reproduces the initiation, propagation, and termination of blood clot formation and its lysis in vitro due to the activation of either the intrinsic or extrinsic pathways. Furthermore, the ABMS was able to simulate the pharmacological effects of two clinically used anticoagulants, warfarin and heparin, as well as the physiological effects of enzyme deficiency/dysfunction, i.e., hemophilia and antithrombin III-heparin binding impairment, on the coagulation system. The results of the model compare favorably with in vitro experimental data under both physiologic and pathophysiologic conditions. Our computational systems biology approach integrates reductionist experimental data into a cohesive model that allows rapid evaluation of the effects of multiple variables. Our ODE and AMBS models offer the ability to generate non-linear responses based on known relationships among variables and in silico modeling of mechanistic biological rules on computer software, respectively. Simulations of normal and disease states as well as effects of therapeutic intervention demonstrate the potential uses of computer simulation. Specifically, models may be applied to hypothesis generation and biological advances, discovery of new diagnostic and therapeutic options, platforms to test novel therapies, and opportunities to predict adverse events during drug development. The ultimate aim of such models is creation of bedside simulators that allow personalized, individual medicine; however, a myriad of opportunities for scientific advancement are opened through in silico experimentation.

The aim of this thesis is to build the foundations of generalized ordinary differ- ential equation theory in metric spaces. While differential equations in metric spaces have been studied before, the chosen approach cannot be extended to in- clude more general types of integral equations. We introduce a definition which combines the added generality of metric spaces with the strength of Kurzweil's generalized ordinary differential equations. Additionally, we present existence and uniqueness theorems which offer new results even in the context of Euclidean spaces.

O objetivo deste trabalho é investigar propriedades qualitativas das soluções de equações diferenciais funcionais com retardamento e impulsos em tempo variável (EDFRs impulsivas) através da teoria de equações diferenciais ordinárias generalizadas (EDOs generalizadas). Nossos principais resultados dizem respeito a estabilidade uniforme, estabilidade uniforme assintótica e estabilidade exponencial da solução trivial de uma determinada classe de EDFRs com impulsos em tempo variável e limitação uniforme de soluções da mesma classe. A fim de obtermos tais resultados para EDFRs com impulsos em tempo variável, estabelecemos novos resultados sobre propriedades qualitativas das soluções de EDOs generalizadas. Assim, portanto, este trabalho contribui para o desenvolvimento de ambas as teorias de EDFRs com impulsos e de EDOs generalizadas. Os resultados novos apresentados neste trabalho estão contidos nos artigos [1], [2] e [3] / The purpose of this work is to investigate qualitative properties of solutions of retarded functional differential equations (RFDEs) with impulse effects acting on variable times using the theory of generalized ordinary differential equations (generalized ODEs). Our main results concern uniform stability, uniform asymptotic stability and exponential stability of the trivial solution of a certain class of RFDEs with variable impulses and uniform boundedness of the solutions of the same class. In order to obtain such results for RFDEs with variable impulses, we establish new results about qualitative properties of solutions of generalized ODEs. In this manner, we contribute with new results not only to the theory of RFDEs with impulses but also to the theory of generalized ODEs. The new results presented in this work are contained in the articles [1], [2] and [3]

O objetivo deste trabalho é investigar propriedades qualitativas das soluções de equações diferenciais funcionais com retardamento e impulsos em tempo variável (EDFRs impulsivas) através da teoria de equações diferenciais ordinárias generalizadas (EDOs generalizadas). Nossos principais resultados dizem respeito a estabilidade uniforme, estabilidade uniforme assintótica e estabilidade exponencial da solução trivial de uma determinada classe de EDFRs com impulsos em tempo variável e limitação uniforme de soluções da mesma classe. A fim de obtermos tais resultados para EDFRs com impulsos em tempo variável, estabelecemos novos resultados sobre propriedades qualitativas das soluções de EDOs generalizadas. Assim, portanto, este trabalho contribui para o desenvolvimento de ambas as teorias de EDFRs com impulsos e de EDOs generalizadas. Os resultados novos apresentados neste trabalho estão contidos nos artigos [1], [2] e [3] / The purpose of this work is to investigate qualitative properties of solutions of retarded functional differential equations (RFDEs) with impulse effects acting on variable times using the theory of generalized ordinary differential equations (generalized ODEs). Our main results concern uniform stability, uniform asymptotic stability and exponential stability of the trivial solution of a certain class of RFDEs with variable impulses and uniform boundedness of the solutions of the same class. In order to obtain such results for RFDEs with variable impulses, we establish new results about qualitative properties of solutions of generalized ODEs. In this manner, we contribute with new results not only to the theory of RFDEs with impulses but also to the theory of generalized ODEs. The new results presented in this work are contained in the articles [1], [2] and [3]

Vancomycin is an antibiotic used for the treatment of systemic infections. It is given intravenously usually every twelve or twenty-four hours. This particular drug has a medium level of boundedness, with approximately fty percent of the drug being free and thus physiologically eective. A physiologically-based pharmacokinetic (PBPK) model was used to better understand the absorption, distribution, and elimination of the drug. Using optimal parameters, the model could be used in the future to test how various factors, such as BMI or excretion levels, might aect the concentration of the antibiotic.

Levofloxacin is in a class of antibiotics known as fluoroquinolones, which treat infections by killing the bacteria that cause them. A physiologically-based pharmacokinetic (PBPK) model was developed to investigate the uptake, distribution, and elimination of Levofloxacin after a single dose. PBPK modeling uses parameters such as body weight, blood flow rates, partition coefficients, organ volumes, and several other parameters in order to model the distribution of a particular drug throughout the body. Levofloxacin is only moderately bound in human blood plasma, and, thus, for the purposes of this paper, linear bonding is incorporated into the model because the free or unbound portion of the drug is the only portion that is considered to be medicinally effective. Parameter estimation is then used to estimate the two unknown parameters given clinical data from literature on the total concentration of Levofloxacin in the blood over time. Once an adequate model is generated, the effects of varying Body Mass Index are tested for the absorption and distribution of Levofloxacin throughout the body.